RebuildStarted event detected on md device /dev/md0
Logging into one of my systems, I saw the following in /proc/mdstat this morning.
Personalities : [raid1]
md0 : active raid1 sdc1[0] sdb1[1]
976559104 blocks [2/2] [UU]
[=========>...........] resync = 45.7% (446333440/976559104) finish=489.8min speed=18037K/sec
Unaware of a power failure or disk failure (these are pretty new disks) I did some digging. Looking through /var/log/daemon.log, I found these entries:
Mar 1 01:06:02 chewbacca mdadm: RebuildStarted event detected on md device /dev/md0
Mar 1 04:49:03 chewbacca mdadm: Rebuild20 event detected on md device /dev/md0
Mar 1 08:43:03 chewbacca mdadm: Rebuild40 event detected on md device /dev/md0
After some googling around, I found that on the first Sunday of every month at 1:06am, an array check (/usr/share/mdadm/checkarray) is run on Debian and Ubuntu systems. (see /etc/cron.d/mdadm)
For some reason it is listed in /proc/mdstat as a rebuild even though it is really a read-only operation to check the health of the array. Whew! Now I can go get ready for church.
Hi altj, I can't tell you how
Hi altj,
I can't tell you how stunned I got after reading this post...
In light of this problem, and since my system crashes everytime this task is ran, do you recommend commenting out this cron task?
Any other advice is welcome ;)
Thanks!
It's not really a problem.
It's not really a problem. Since it's just a verification process, you can comment out this cron task. You may want to run it manually every once in a while just to make sure your RAID is healthy.
Thanks for the reply
Thanks for the reply altj,
It's a problem to me because everytime the check runs my server freezes and shows a red warning light :P
I'm using mdadm 2.6.3 and I already tried to coment out this task, setting the AUTOCHECK parameter on /etc/default/mdadm to false. Strangelly and despite no rebuild event is shown on the daemon log, by doing this, the system freezes anyday at 1.06 am, but this time no red warning light is show.
I'me shure that the disks are ok, because the server works perfectly and I don't have any (F) notification associated with any device of each array at /proc/mdstat.
I would really like to come up with a solution that doesn'e require me to disable the checkarry, and have severall ideias that I will teste these days.
Thanks once again!
Post new comment